English
 
User Manual Privacy Policy Disclaimer Contact us
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT

Released

Poster

Causal inference conditions reliability-weighted integration of audiovisual spatial signals

MPS-Authors
/persons/resource/persons84450

Rohe,  T
Max Planck Institute for Biological Cybernetics, Max Planck Society;
Research Group Cognitive Neuroimaging, Max Planck Institute for Biological Cybernetics, Max Planck Society;

/persons/resource/persons84112

Noppeney,  U
Research Group Cognitive Neuroimaging, Max Planck Institute for Biological Cybernetics, Max Planck Society;
Max Planck Institute for Biological Cybernetics, Max Planck Society;

External Ressource

Link
(Abstract)

Fulltext (public)
There are no public fulltexts stored in PuRe
Supplementary Material (public)
There is no public supplementary material available
Citation

Rohe, T., & Noppeney, U. (2013). Causal inference conditions reliability-weighted integration of audiovisual spatial signals. Poster presented at Bernstein Conference 2013, Tübingen, Germany.


Cite as: http://hdl.handle.net/21.11116/0000-0001-4E47-9
Abstract
To form coherent and reliable multisensory percepts of the environment, human observers have to segregate multisensory signals caused by independent sources but integrate those from a common source. Models of causal inferences (Kording et al., 2007) predict the inference of a common cause if the signals are close in space and time. Further, models of optimal reliability-weighted integration predict that multisensory signals are weighed proportional to their relative reliability in order to maximize the reliability of the integrated percept (Ernst Banks, 2002). To probe models of causal inference and reliability-weighted integration, we presented subjects (N = 26) with audiovisual spatial cues and manipulated spatial disparity and visual reliability. Subjects were required to selectively localize the auditory cues and to judge the spatial unity of the cues. Indices of audiovisual spatial integration showed that audiovisual spatial cues were weighted proportional to visual reliability, but only if a common cause was inferred. Likewise, localization reliability increased with visual reliability in case of a common-cause inference. Computational models incorporating causal inferences and reliability-weighted integration provided superior fit to auditory-localization data compared to models implementing only reliability-weighted integration. The results suggest that reliability-weighed integration is conditioned on the outcome of the causal inference.